1 November 2019
Caroline Uhler
Department of Electrical Engineering and Computer Science and
Institute for Data, Systems and Society
MIT
Recent evidence suggests that the spatial organization of the genome represents an important epigenetic regulator of expression and alterations thereof are associated with various diseases. In this talk, I will analyze the link between the 3D genome organization and gene regulation at the single-cell level by combining sequencing data (drop-seq, perturb-seq, Hi-C) with imaging data (3D FISH) and modeling (causal inference, geometric packing models, and autoencoders/deep learning). In particular, I will propose a new chromosome packing model that links the mechanical state of a cell with gene expression via functional gene clusters in the chromosome intermingling regions. In addition, I will discuss causal inference algorithms for learning gene regulatory networks by taking advantage of the high-throughput interventional single-cell RNA-seq data that is currently being produced. Furthermore, I will discuss approaches for integrating different data modalities such as sequencing or imaging via autoencoders. Finally, I will present a machine learning pipeline for detecting changes in genome packing at the single-cell level as a biomarker for early cancer diagnostics. Collectively, our studies provide important insights into the spatial control of genetic information in health and disease.